iFLYTEK Prepares Major AI Model Launch on Domestic Compute, Targeting Leadership in Code and Value
Leading AI company iFLYTEK is finalizing a new flagship general-purpose model, uniquely trained and powered entirely by China's domestic computing infrastructure. The move signals a significant step in the country's pursuit of self-reliant AI technology.
Aiming for Top-Tier Performance Metrics
According to President Wu Xiaoru, the upcoming model is designed to excel in two critical areas: code-related capabilities and overall token cost-efficiency. Its combined performance is projected to place it among the very best domestic models currently available.
Maintaining Leadership in the Domestic Compute Arena
This model is distinguished by its complete reliance on nationally-sourced computing power for training. Within this specific category of models—those built on domestic hardware—iFLYTEK's offering is expected to maintain leading positions across key benchmarks, reinforcing the viability of the homegrown tech stack.
On-Device Deployment Strategy
Beyond cloud services, iFLYTEK has concrete plans for edge deployment. The model is slated for integration into real-world applications such as intelligent vehicle systems and robotics, bridging the gap between AI algorithms and physical interaction.
A staged version of the model is scheduled for release by the end of August this year, offering a first look at its capabilities and setting a timeline for its broader rollout.